smids_3x_deit_tiny_sgd_001_fold5
This model is a fine-tuned version of facebook/deit-tiny-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.2846
- Accuracy: 0.8767
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.8735 | 1.0 | 225 | 0.9170 | 0.545 |
| 0.6877 | 2.0 | 450 | 0.6945 | 0.7167 |
| 0.5875 | 3.0 | 675 | 0.5531 | 0.79 |
| 0.4761 | 4.0 | 900 | 0.4755 | 0.82 |
| 0.4339 | 5.0 | 1125 | 0.4319 | 0.83 |
| 0.3839 | 6.0 | 1350 | 0.4018 | 0.8483 |
| 0.4408 | 7.0 | 1575 | 0.3811 | 0.85 |
| 0.4281 | 8.0 | 1800 | 0.3653 | 0.8583 |
| 0.3652 | 9.0 | 2025 | 0.3551 | 0.8533 |
| 0.324 | 10.0 | 2250 | 0.3480 | 0.8567 |
| 0.3571 | 11.0 | 2475 | 0.3391 | 0.86 |
| 0.3339 | 12.0 | 2700 | 0.3292 | 0.86 |
| 0.343 | 13.0 | 2925 | 0.3232 | 0.865 |
| 0.3099 | 14.0 | 3150 | 0.3188 | 0.8633 |
| 0.2636 | 15.0 | 3375 | 0.3160 | 0.87 |
| 0.2725 | 16.0 | 3600 | 0.3109 | 0.8633 |
| 0.2598 | 17.0 | 3825 | 0.3041 | 0.8717 |
| 0.2377 | 18.0 | 4050 | 0.3051 | 0.8683 |
| 0.2636 | 19.0 | 4275 | 0.2990 | 0.8683 |
| 0.2944 | 20.0 | 4500 | 0.2992 | 0.8733 |
| 0.2247 | 21.0 | 4725 | 0.2983 | 0.8733 |
| 0.2126 | 22.0 | 4950 | 0.2963 | 0.8733 |
| 0.2221 | 23.0 | 5175 | 0.2922 | 0.8783 |
| 0.2198 | 24.0 | 5400 | 0.2918 | 0.8717 |
| 0.2574 | 25.0 | 5625 | 0.2955 | 0.88 |
| 0.2932 | 26.0 | 5850 | 0.2903 | 0.88 |
| 0.2755 | 27.0 | 6075 | 0.2866 | 0.8767 |
| 0.2735 | 28.0 | 6300 | 0.2890 | 0.8783 |
| 0.2207 | 29.0 | 6525 | 0.2874 | 0.88 |
| 0.1879 | 30.0 | 6750 | 0.2869 | 0.875 |
| 0.1763 | 31.0 | 6975 | 0.2867 | 0.88 |
| 0.2308 | 32.0 | 7200 | 0.2871 | 0.8733 |
| 0.1914 | 33.0 | 7425 | 0.2850 | 0.8767 |
| 0.1699 | 34.0 | 7650 | 0.2866 | 0.875 |
| 0.1804 | 35.0 | 7875 | 0.2842 | 0.88 |
| 0.182 | 36.0 | 8100 | 0.2861 | 0.8783 |
| 0.2385 | 37.0 | 8325 | 0.2854 | 0.8783 |
| 0.1637 | 38.0 | 8550 | 0.2879 | 0.8783 |
| 0.1554 | 39.0 | 8775 | 0.2876 | 0.8767 |
| 0.2151 | 40.0 | 9000 | 0.2856 | 0.8783 |
| 0.1931 | 41.0 | 9225 | 0.2849 | 0.8783 |
| 0.1711 | 42.0 | 9450 | 0.2846 | 0.8783 |
| 0.2357 | 43.0 | 9675 | 0.2864 | 0.8783 |
| 0.202 | 44.0 | 9900 | 0.2845 | 0.88 |
| 0.1905 | 45.0 | 10125 | 0.2855 | 0.88 |
| 0.1822 | 46.0 | 10350 | 0.2847 | 0.8767 |
| 0.2034 | 47.0 | 10575 | 0.2849 | 0.8767 |
| 0.1793 | 48.0 | 10800 | 0.2851 | 0.8767 |
| 0.2049 | 49.0 | 11025 | 0.2846 | 0.8767 |
| 0.168 | 50.0 | 11250 | 0.2846 | 0.8767 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2